External validation and updating of the infusion rate individualization of soybean oil-based intravenous lipid emulsion: A descriptive cohort study.

Abstract

Safe and efficient provision of intravenous lipid emulsion (ILE) requires a strategy to individualize infusion rates. Estimating the maximum acceptable infusion rate (MaxInfRate) of soybean oil-based ILE (SO-ILE) in individuals by using a triglyceride (TG) kinetic model was reported to be feasible. In this study, we aimed to externally validate and, if needed, update the MaxInfRate estimation. The maximum TG concentration (TGmax) in patients receiving SO-ILE at MaxInfRate was evaluated to determine if it met the definition of being <400 mg/dl for 90th percentile of patients. The TG kinetic model was evaluated through prediction performance checks and was subsequently updated using the data set of both the previous model development and present validation studies. Out of 83 patients, 74 had TGmax <400 mg/dl, corresponding to a probability of 89.2% (95% CI, 81.9%-95.2%), and the 90th percentile of TGmax was 400 mg/dl (95% CI, 328-490 mg/dl), closely aligned with the theoretical values. However, the individual TGmax values were biased by the infusion rate because the covariate effects were overestimated in the TG kinetic model, requiring a minor revision. The updated MaxInfRate with the combined data set showed unbiased and more accurate predictions. The MaxInfRate was validated in external inpatients and updated with all available data. MaxInfRate estimation for individuals could be an option for the safe and efficient provision of SO-ILE.